In the previous chapters, we have learnt about the tools and techniques of Quantitative Risk Analysis that come under the Data Gathering and Representation Techniques sub-group. In this chapter, we are going to take a look at Sensitivity Analysis that falls under the Quantitative Risk Analysis & Modeling Techniques sub-group.
Purpose of Sensitivity Analysis:
The purpose of sensitivity analysis is to determine which risks have the highest potential impact on project objectives. Our goal is to single out those important risks so that we can respond in a more effective manner. During our analysis, we will find out things like:
a. How imp certain elements are to the project?
b. Which variables require special attention?
The Use of Sensitivity Analysis
Any analysis we take up should have some direct or indirect use, otherwise, what is the point of taking it up? Similarly, Sensitivity analysis can help us gather information to back up our recommendations. Sensitivity Analysis shows us a range of outcomes for a risk event. It also helps us identify those risks that can have the greatest effect on the project plan and the overall project objectives.
Once we know which risk would impact us more, we can use the Sensitivity Analysis as the information backing up our recommendation to give additional consideration for that risk.
How Sensitivity Analysis is performed:
We measure the effects of a project element on the project objective, when all the other elements are held at their baseline values. By modifying one element and keeping the others static, we are trying to determine the level of uncertainty each element could pose to our project objectives.
Information required to perform Sensitivity Analysis:
We will be using all the quantitative information gathered for the risks up until now, along with other information from the Project Management Plan. Remember that these risks are expected to have a significant impact on the project objectives like cost, time, quality etc. So, the corresponding plans from the Project Management Plan too may be consulted when the analysis is performed.
Another point to note here is that, Sensitivity analysis can be done on a risk even if it does not fall under the Important or High Impact category. But, in real life, we may not have the time or the resources to take up such analysis on lower impact/priority risks.
What happens after Sensitivity Analysis?
We establish a range of variation for each of the risk event and also determine a level of acceptance. Our focus is on changing a single element only. This is also called “What-If Scenarios” wherein we understand what happens if a particular element that could impact the project objective is altered while all others remain stable. We typically play around with the variables and see what happens. Let me repeat, during normal sensitivity analysis, only one element/variable is changed at any given time.
If you are thinking, why should I change only one variable, why can’t I change more than I at the same time, then fear not? Whatever you thought is perfectly valid and possible. If you do that, it’s perfectly fine but, it just won’t be called Sensitivity Analysis. It is called “Design of Experiments”
Design of experiments is a statistical method to find the factors that may influence specific variables that may affect the specific outcomes of our project. We determine the combined effect of uncertainty as well as interaction between the various factors.
Once all your analysis is complete, you need to represent your findings in some form so that the other people in your team as well as your management can understand it. Isn’t it? This final representation is called a “Tornado Diagram”
Tornado Diagrams are a very common way of displaying results of sensitivity analysis. They compare the importance of variables that have a higher degree of uncertainty to the more stable variables. As you might remember from the previous paragraphs, during sensitivity analysis, one variables impact on the project objectives is understood while all other variables are set at the baseline.
A sample tornado diagram:
As you can see, the greater the effect of a variable, the higher up it will feature on the diagram. This means that we should focus on the elements that are higher up in the image.
When you say that a particular risk would have a higher impact on the project and back it up with the sensitivity analysis and tornado diagram, it would be easier for the people higher up the org hierarchy to understand your recommendation.
Typically we use tornado diagrams to represent impact on cost, time or quality objectives.
Prev: Discrete Distributions
Next: Expected Monetary Value Analysis